This report describes how to construct time series for the price of cocaine using data from the Drug Enforcement Administration's System to Retrieve Information from Drug Evidence, a database that includes records of prices paid by undercover agents for individual purchases. Central to this process is the task of standardizing data for transaction size and purity. Prior efforts in this area are reviewed and their treatment of purity found wanting. This report suggests that because quality control is difficult for illicit products, price is governed more by the expected purity than by the actual purity of the product. Using this concept, price series are constructed for the gram, ounce, and kilogram level in a variety of locations. Analysis of these series reveals that significant price differences exist between cities, even at the wholesale level; these differences do not necessarily dissipate over time; and the ratio of prices at different market levels has remained remarkably constant over time. This last result is consistent with the hypothesis that price increases at one level are passed through to lower levels on a percentage basis (i.e., according to a multiplicative model) rather than a dollar-for-dollar basis (i.e., according to an additive model).